Bundle of miscellaneous ia64 fixes for 3.10 merge window.

-----BEGIN PGP SIGNATURE-----
 Version: GnuPG v1.4.11 (GNU/Linux)
 
 iQIcBAABAgAGBQJReu6gAAoJEKurIx+X31iB9aoQAItXGa7kGdtPXBBgnQVx7Q/T
 B9LbO3qVAnEEQ7x73C8NUAiDfT8KKwDbHZDyMVqlyZSfckAtNcZikXGuWZfHIktc
 7Xg1ogLEAPDiUCKLXV/zGMCwn08ew+QNycwJrxjSehXfYZFdj4JTCIDswJ1RUNCT
 D0U+afhsWuZRH6yZOEFzap2ZwX+o2m4jSa0tHO0eiXAcOTTF8Nxy8mDZKBJ7kNrg
 2WxYvNjTnNR3csRRmdRnQEKHtvI/snGRkMS4UoOKDN3+IZo8OsKNbWfFhysFoYVq
 gGXSTI8lqG7j+POCQdp2k4XvmoOhVyll79Bo0bYOSMO0gkqUp86+2i/IWx7L/njS
 4wbi5nMSzhX4t4GBpEo99NV76KLKaSwjAUVTWhbM7L40NFjSz3rXm5qfbh8Y+k/v
 ngi41yx1nOkOYJCSRvjjFAUGnHNE/k/7aSGb+Nba13RnJlJ36tpnifszqxhOax7+
 BmXn5S2lOLh1wrM/nBOKmbfrk/+N9QI5MhVoyA/Fg3SCzl96T7gVVEAfCfTmC2vQ
 r/6/djRVfIhiLcHhdV3ev6TC41AynMAAuOuClU+lw2ZO8kOn9fHoKV9Mz2J1flQq
 gJEfIuKToYbizDDXa+wf/k4G5GD77D7AKlhlZw6XcKJImCWktCzzlwOqwMIj27sj
 F00m4RtrlFHUBTT+tvc2
 =Y2HM
 -----END PGP SIGNATURE-----

Merge tag 'please-pull-misc-3.10' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux

Pull ia64 fixes from Tony Luck:
 "Bundle of miscellaneous ia64 fixes for 3.10 merge window."

* tag 'please-pull-misc-3.10' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux:
  Add size restriction to the kdump documentation
  Fix example error_injection_tool
  Fix build error for numa_clear_node() under IA64
  Fix initialization of CMCI/CMCP interrupts
  Change "select DMAR" to "select INTEL_IOMMU"
  Wrong asm register contraints in the kvm implementation
  Wrong asm register contraints in the futex implementation
  Remove cast for kmalloc return value
  Fix kexec oops when iosapic was removed
  iosapic: fix a minor typo in comments
  Add WB/UC check for early_ioremap
  Fix broken fsys_getppid()
  tiocx: check retval from bus_register()

Documentation/ia64/err_inject.txt

@@ -882,7 +882,7 @@ int err_inj()
 			cpu=parameters[i].cpu;
 			k = cpu%64;
 			j = cpu/64;
-			mask[j]=1<<k;
+			mask[j] = 1UL << k;
 
 			if (sched_setaffinity(0, MASK_SIZE*8, mask)==-1) {
 				perror("Error sched_setaffinity:");

Documentation/kdump/kdump.txt

@@ -297,6 +297,7 @@ Boot into System Kernel
    On ia64, 256M@256M is a generous value that typically works.
    The region may be automatically placed on ia64, see the
    dump-capture kernel config option notes above.
+   If use sparse memory, the size should be rounded to GRANULE boundaries.
 
    On s390x, typically use "crashkernel=xxM". The value of xx is dependent
    on the memory consumption of the kdump system. In general this is not

arch/ia64/Kconfig

@@ -187,7 +187,7 @@ config IA64_DIG
 
 config IA64_DIG_VTD
 	bool "DIG+Intel+IOMMU"
-	select DMAR
+	select INTEL_IOMMU
 	select PCI_MSI
 
 config IA64_HP_ZX1

arch/ia64/include/asm/futex.h

@@ -106,16 +106,15 @@ futex_atomic_cmpxchg_inatomic(u32 *uval, u32 __user *uaddr,
 		return -EFAULT;
 
 	{
-		register unsigned long r8 __asm ("r8");
+		register unsigned long r8 __asm ("r8") = 0;
 		unsigned long prev;
 		__asm__ __volatile__(
 			"	mf;;					\n"
-			"	mov %0=r0				\n"
 			"	mov ar.ccv=%4;;				\n"
 			"[1:]	cmpxchg4.acq %1=[%2],%3,ar.ccv		\n"
 			"	.xdata4 \"__ex_table\", 1b-., 2f-.	\n"
 			"[2:]"
-			: "=r" (r8), "=r" (prev)
+			: "+r" (r8), "=&r" (prev)
 			: "r" (uaddr), "r" (newval),
 			  "rO" ((long) (unsigned) oldval)
 			: "memory");

arch/ia64/include/asm/mca.h

@@ -143,6 +143,7 @@ extern unsigned long __per_cpu_mca[NR_CPUS];
 extern int cpe_vector;
 extern int ia64_cpe_irq;
 extern void ia64_mca_init(void);
+extern void ia64_mca_irq_init(void);
 extern void ia64_mca_cpu_init(void *);
 extern void ia64_os_mca_dispatch(void);
 extern void ia64_os_mca_dispatch_end(void);

arch/ia64/include/asm/numa.h

@@ -67,14 +67,13 @@ extern int paddr_to_nid(unsigned long paddr);
 
 extern void map_cpu_to_node(int cpu, int nid);
 extern void unmap_cpu_from_node(int cpu, int nid);
-
+extern void numa_clear_node(int cpu);
 
 #else /* !CONFIG_NUMA */
 #define map_cpu_to_node(cpu, nid)	do{}while(0)
 #define unmap_cpu_from_node(cpu, nid)	do{}while(0)
-
 #define paddr_to_nid(addr)	0
-
+#define numa_clear_node(cpu)	do { } while (0)
 #endif /* CONFIG_NUMA */
 
 #endif /* _ASM_IA64_NUMA_H */

arch/ia64/kernel/fsys.S

@@ -90,53 +90,6 @@ ENTRY(fsys_getpid)
 	FSYS_RETURN
 END(fsys_getpid)
 
-ENTRY(fsys_getppid)
-	.prologue
-	.altrp b6
-	.body
-	add r17=IA64_TASK_GROUP_LEADER_OFFSET,r16
-	;;
-	ld8 r17=[r17]				// r17 = current->group_leader
-	add r9=TI_FLAGS+IA64_TASK_SIZE,r16
-	;;
-
-	ld4 r9=[r9]
-	add r17=IA64_TASK_REAL_PARENT_OFFSET,r17 // r17 = &current->group_leader->real_parent
-	;;
-	and r9=TIF_ALLWORK_MASK,r9
-
-1:	ld8 r18=[r17]				// r18 = current->group_leader->real_parent
-	;;
-	cmp.ne p8,p0=0,r9
-	add r8=IA64_TASK_TGID_OFFSET,r18	// r8 = &current->group_leader->real_parent->tgid
-	;;
-
-	/*
-	 * The .acq is needed to ensure that the read of tgid has returned its data before
-	 * we re-check "real_parent".
-	 */
-	ld4.acq r8=[r8]				// r8 = current->group_leader->real_parent->tgid
-#ifdef CONFIG_SMP
-	/*
-	 * Re-read current->group_leader->real_parent.
-	 */
-	ld8 r19=[r17]				// r19 = current->group_leader->real_parent
-(p8)	br.spnt.many fsys_fallback_syscall
-	;;
-	cmp.ne p6,p0=r18,r19			// did real_parent change?
-	mov r19=0			// i must not leak kernel bits...
-(p6)	br.cond.spnt.few 1b			// yes -> redo the read of tgid and the check
-	;;
-	mov r17=0			// i must not leak kernel bits...
-	mov r18=0			// i must not leak kernel bits...
-#else
-	mov r17=0			// i must not leak kernel bits...
-	mov r18=0			// i must not leak kernel bits...
-	mov r19=0			// i must not leak kernel bits...
-#endif
-	FSYS_RETURN
-END(fsys_getppid)
-
 ENTRY(fsys_set_tid_address)
 	.prologue
 	.altrp b6
@@ -614,7 +567,7 @@ paravirt_fsyscall_table:
 	data8 0				// chown
 	data8 0				// lseek		// 1040
 	data8 fsys_getpid		// getpid
-	data8 fsys_getppid		// getppid
+	data8 0				// getppid
 	data8 0				// mount
 	data8 0				// umount
 	data8 0				// setuid		// 1045

arch/ia64/kernel/iosapic.c

@@ -76,7 +76,7 @@
  *	PCI pin -> global system interrupt (GSI) -> IA-64 vector <-> IRQ
  *
  * Note: The term "IRQ" is loosely used everywhere in Linux kernel to
- * describeinterrupts.  Now we use "IRQ" only for Linux IRQ's.  ISA IRQ
+ * describe interrupts.  Now we use "IRQ" only for Linux IRQ's.  ISA IRQ
  * (isa_irq) is the only exception in this source code.
  */
 
@@ -1010,6 +1010,26 @@ iosapic_check_gsi_range (unsigned int gsi_base, unsigned int ver)
 	return 0;
 }
 
+static int
+iosapic_delete_rte(unsigned int irq, unsigned int gsi)
+{
+	struct iosapic_rte_info *rte, *temp;
+
+	list_for_each_entry_safe(rte, temp, &iosapic_intr_info[irq].rtes,
+								rte_list) {
+		if (rte->iosapic->gsi_base + rte->rte_index == gsi) {
+			if (rte->refcnt)
+				return -EBUSY;
+
+			list_del(&rte->rte_list);
+			kfree(rte);
+			return 0;
+		}
+	}
+
+	return -EINVAL;
+}
+
 int iosapic_init(unsigned long phys_addr, unsigned int gsi_base)
 {
 	int num_rte, err, index;
@@ -1069,7 +1089,7 @@ int iosapic_init(unsigned long phys_addr, unsigned int gsi_base)
 
 int iosapic_remove(unsigned int gsi_base)
 {
-	int index, err = 0;
+	int i, irq, index, err = 0;
 	unsigned long flags;
 
 	spin_lock_irqsave(&iosapic_lock, flags);
@@ -1087,6 +1107,16 @@ int iosapic_remove(unsigned int gsi_base)
 		goto out;
 	}
 
+	for (i = gsi_base; i < gsi_base + iosapic_lists[index].num_rte; i++) {
+		irq = __gsi_to_irq(i);
+		if (irq < 0)
+			continue;
+
+		err = iosapic_delete_rte(irq, i);
+		if (err)
+			goto out;
+	}
+
 	iounmap(iosapic_lists[index].addr);
 	iosapic_free(index);
  out:

arch/ia64/kernel/irq.c

@@ -23,6 +23,8 @@
 #include <linux/interrupt.h>
 #include <linux/kernel_stat.h>
 
+#include <asm/mca.h>
+
 /*
  * 'what should we do if we get a hw irq event on an illegal vector'.
  * each architecture has to answer this themselves.
@@ -83,6 +85,12 @@ bool is_affinity_mask_valid(const struct cpumask *cpumask)
 
 #endif /* CONFIG_SMP */
 
+int __init arch_early_irq_init(void)
+{
+	ia64_mca_irq_init();
+	return 0;
+}
+
 #ifdef CONFIG_HOTPLUG_CPU
 unsigned int vectors_in_migration[NR_IRQS];
 

arch/ia64/kernel/mca.c

@@ -2074,22 +2074,16 @@ ia64_mca_init(void)
 	printk(KERN_INFO "MCA related initialization done\n");
 }
 
-/*
- * ia64_mca_late_init
- *
- *	Opportunity to setup things that require initialization later
- *	than ia64_mca_init.  Setup a timer to poll for CPEs if the
- *	platform doesn't support an interrupt driven mechanism.
- *
- *  Inputs  :   None
- *  Outputs :   Status
- */
-static int __init
-ia64_mca_late_init(void)
-{
-	if (!mca_init)
-		return 0;
 
+/*
+ * These pieces cannot be done in ia64_mca_init() because it is called before
+ * early_irq_init() which would wipe out our percpu irq registrations. But we
+ * cannot leave them until ia64_mca_late_init() because by then all the other
+ * processors have been brought online and have set their own CMC vectors to
+ * point at a non-existant action. Called from arch_early_irq_init().
+ */
+void __init ia64_mca_irq_init(void)
+{
 	/*
 	 *  Configure the CMCI/P vector and handler. Interrupts for CMC are
 	 *  per-processor, so AP CMC interrupts are setup in smp_callin() (smpboot.c).
@@ -2108,6 +2102,23 @@ ia64_mca_late_init(void)
 	/* Setup the CPEI/P handler */
 	register_percpu_irq(IA64_CPEP_VECTOR, &mca_cpep_irqaction);
 #endif
+}
+
+/*
+ * ia64_mca_late_init
+ *
+ *	Opportunity to setup things that require initialization later
+ *	than ia64_mca_init.  Setup a timer to poll for CPEs if the
+ *	platform doesn't support an interrupt driven mechanism.
+ *
+ *  Inputs  :   None
+ *  Outputs :   Status
+ */
+static int __init
+ia64_mca_late_init(void)
+{
+	if (!mca_init)
+		return 0;
 
 	register_hotcpu_notifier(&mca_cpu_notifier);
 

arch/ia64/kernel/mca_drv.c

@@ -349,7 +349,7 @@ init_record_index_pools(void)
 
 	/* - 3 - */
 	slidx_pool.max_idx = (rec_max_size/sect_min_size) * 2 + 1;
-	slidx_pool.buffer = (slidx_list_t *)
+	slidx_pool.buffer =
 		kmalloc(slidx_pool.max_idx * sizeof(slidx_list_t), GFP_KERNEL);
 
 	return slidx_pool.buffer ? 0 : -ENOMEM;

arch/ia64/kvm/vtlb.c

@@ -256,7 +256,7 @@ u64 guest_vhpt_lookup(u64 iha, u64 *pte)
 			"srlz.d;;"
 			"ssm psr.i;;"
 			"srlz.d;;"
-			: "=r"(ret) : "r"(iha), "r"(pte):"memory");
+			: "=&r"(ret) : "r"(iha), "r"(pte) : "memory");
 
 	return ret;
 }

arch/ia64/mm/ioremap.c

@@ -16,7 +16,7 @@
 #include <asm/meminit.h>
 
 static inline void __iomem *
-__ioremap (unsigned long phys_addr)
+__ioremap_uc(unsigned long phys_addr)
 {
 	return (void __iomem *) (__IA64_UNCACHED_OFFSET | phys_addr);
 }
@@ -24,7 +24,11 @@ __ioremap (unsigned long phys_addr)
 void __iomem *
 early_ioremap (unsigned long phys_addr, unsigned long size)
 {
-	return __ioremap(phys_addr);
+	u64 attr;
+	attr = kern_mem_attribute(phys_addr, size);
+	if (attr & EFI_MEMORY_WB)
+		return (void __iomem *) phys_to_virt(phys_addr);
+	return __ioremap_uc(phys_addr);
 }
 
 void __iomem *
@@ -47,7 +51,7 @@ ioremap (unsigned long phys_addr, unsigned long size)
 	if (attr & EFI_MEMORY_WB)
 		return (void __iomem *) phys_to_virt(phys_addr);
 	else if (attr & EFI_MEMORY_UC)
-		return __ioremap(phys_addr);
+		return __ioremap_uc(phys_addr);
 
 	/*
 	 * Some chipsets don't support UC access to memory.  If
@@ -93,7 +97,7 @@ ioremap (unsigned long phys_addr, unsigned long size)
 		return (void __iomem *) (offset + (char __iomem *)addr);
 	}
 
-	return __ioremap(phys_addr);
+	return __ioremap_uc(phys_addr);
 }
 EXPORT_SYMBOL(ioremap);
 
@@ -103,7 +107,7 @@ ioremap_nocache (unsigned long phys_addr, unsigned long size)
 	if (kern_mem_attribute(phys_addr, size) & EFI_MEMORY_WB)
 		return NULL;
 
-	return __ioremap(phys_addr);
+	return __ioremap_uc(phys_addr);
 }
 EXPORT_SYMBOL(ioremap_nocache);
 

arch/ia64/mm/numa.c

@@ -73,6 +73,11 @@ int __meminit __early_pfn_to_nid(unsigned long pfn)
 	return -1;
 }
 
+void __cpuinit numa_clear_node(int cpu)
+{
+	unmap_cpu_from_node(cpu, NUMA_NO_NODE);
+}
+
 #ifdef CONFIG_MEMORY_HOTPLUG
 /*
  *  SRAT information is stored in node_memblk[], then we can use SRAT

arch/ia64/sn/kernel/tiocx.c

@@ -490,11 +490,14 @@ static int __init tiocx_init(void)
 {
 	cnodeid_t cnodeid;
 	int found_tiocx_device = 0;
+	int err;
 
 	if (!ia64_platform_is("sn2"))
 		return 0;
 
-	bus_register(&tiocx_bus_type);
+	err = bus_register(&tiocx_bus_type);
+	if (err)
+		return err;
 
 	for (cnodeid = 0; cnodeid < num_cnodes; cnodeid++) {
 		nasid_t nasid;